Evolution and diversification of the momilactone biosynthetic gene cluster in the genus Oryza

Abstract

Introduction

Biosynthetic gene clusters and their evolution

With more and more plant reference genome assemblies becoming available, biosynthetic gene clusters (BGCs), i.e., the co-localization of often phylogenetically unrelated genes that participate in the same biosynthetic cascade of specialized metabolites, have emerged as a common feature of genomic organization in plants (Polturak et al., 2022b). BGCs are postulated to confer evolutionary advantages because they facilitate coordinated gene expression, enable the reliable coinheritance of genes involved in the same metabolic pathway (thereby preventing the accumulation of toxic intermediates), or facilitate the formation of metabolons (Nützmann et al., 2016). However, the mechanisms by which such nonorthologous genes become localized in the same genomic region and act in the same biosynthetic pathway are still poorly understood. Currently, the most common model proposes that they have formed through a series of events that is driven by both positive- and negative-selection pressure, starting with gene duplication, followed by neofunctionalization, and ultimately relocation. In some cases, this process appears to have been mediated by transposable elements (Polturak et al., 2022b; Smit & Lichman, 2022).

Biological functions of rice phytoalexins, labdane-related diterpenoids and momilactones

Phytoalexins are low-molecular-mass specialized plant metabolites that are often produced under biotic and abiotic stress conditions (Ahuja et al., 2012). In rice (Oryza sativa), the major phytoalexins are a group of labdane-related diterpenoids (reviewed in Toyomasu et al., 2020), which derive from the cyclization of geranylgeranyl diphosphate (GGPP) into ent, syn, or normal stereoisomers of copalyl diphosphate (CDP) by the class II diterpene synthases Copalyl Diphosphate Synthases (CPSs). The biosynthesis of these metabolites has evolved from that of gibberellins (GAs), ent labdane-related diterpenoids themselves, through duplication and neofunctionalization of core biosynthetic enzymes (Zi et al., 2014). Several ent and syn (but not normal) rice labdane-related diterpenoids have been identified, including momilactones A and B, phytocassanes A to F, and oryzalexins (A to F, and S) (Zi et al., 2014; Toyomasu et al., 2020). Notably, momilactone A and, more prominently, momilactone B have a strong allelopathic activity, that is they inhibit the germination and growth of nearby plants upon being released by the rice plants into the soil (Kato et al., 1973; Kato-Noguchi et al., 2010; Serra Serra et al., 2021). Both compounds accumulate in rice husks but are also exuded from the roots (Kato-Noguchi & Ino, 2003; Kato-Noguchi et al., 2010).

Biosynthesis of momilactones and clustering of momilactone genes

Momilactone biosynthesis (Fig. 1a) starts with the cyclization of GGPP into syn-copalyl diphosphate (syn-CDP), catalysed by the Copalyl Diphosphate Synthase 4 (CPS4) (Otomo et al., 2004b; Xu et al., 2004). syn-CDP is further cyclized into 9βH-pimara-7,15-diene (also known as syn-pimaradiene in the literature) by ent-kaurene synthase-like 4 (KSL4), a class I diterpene synthase (Otomo et al., 2004a; Wilderman et al., 2004). Because syn-CDP is also a substrate for oryzalexin S biosynthesis, the KSL4-mediated cyclization is the first truly dedicated step towards momilactone production (Tamogani et al., 1993). 9βH-pimara-7,15-diene undergoes several oxidation steps, catalysed first by cytochrome P450 (CYP) monooxygenases CYP76M8 and CYP99A3, followed by the short-chain alcohol dehydrogenase MOMILACTONE A SYNTHASE (MAS) and CYP701A8, to yield momilactone A (Fig. 1a) (De La Peña & Sattely, 2021; Kitaoka et al., 2021). CYP76M14 catalyses the final hydroxylation of C20, leading to spontaneous closure of the hemi-acetal ring and forming momilactone B (De La Peña & Sattely, 2021). Among the biosynthetic genes, CPS4, KSL4, and the paralogs CYP99A2/3 and MAS1/2 are co-localized on chromosome 4 in a momilactone biosynthetic gene cluster (MBGC) (Shimura et al., 2007; Miyamoto et al., 2016). However, not all genes required for momilactone biosynthesis are contained in the MBGC: CYP76M8 is located on chromosome 2 and is part of another BGC that is required for phytocassane and oryzalexin production (Okada, 2011; Kitaoka et al., 2021), while CYP701A8 and CYP76M14 are located on chromosomes 6 and 1, respectively (De La Peña & Sattely, 2021).

Fig. 1

Open in figure viewerPowerPoint

The momilactone biosynthetic gene cluster in the Oryza genus. (a) Simplified representation of the momilactone biosynthetic pathway, adapted from De La Peña & Sattely (2021) and Kitaoka et al. (2021). (b) Phylogenetic relationship between the different Oryza species and sub-genomes included in this study, the outgroup Leersia perrieri, and their respective MBGCs. The species tree represents the maximum-likelihood tree inferred from a concatenated multiple-sequence alignment of single-copy orthologues using 4069 orthogroups with a minimum of 100.0% of species having single-copy genes in any orthogroup. Coloured block arrows represent genes. Scaffold and positional information on all genes is provided in Table S2. (c) Microsynteny between the genomic regions containing the MBGC from Oryza sativa, Oryza punctata, Oryza officinalis, Oryza alta, Oryza coarctata, Oryza brachyantha, and L. perrieri. In (c), grey lines connect the respective orthologues. ‘Nonrelated’ refers to genes that do not seem to be functionally related to the main biosynthetic cascade of the cluster.

Evolution of the momilactone biosynthetic gene cluster in Oryza

The Oryza genus (belonging to the Poaceae family) consists of 27 known species with 11 different genome types (classified based on cytogenetics and genetic hybridization studies): 6 diploids (n = 12; AA, BB, CC, EE, FF and GG) and 5 allotetraploids (n = 24; BBCC, CCDD, HHJJ, HHKK and KKLL) (Ge et al., 1999; Lu et al., 2009). The MBGC was established in Oryza before the domestication of rice (O. sativa; AA) (Miyamoto et al., 2016). It is highly conserved among AA and BB genome type Oryza species, while only a partial cluster exists in the early-divergent Oryza species O. brachyantha (FF), which harbours only two clustered CYP99A2/3 paralogs (Miyamoto et al., 2016). Because the MBGC is incomplete in O. brachyantha, it likely evolved before the divergence of the BB lineage in Oryza (Miyamoto et al., 2016). Due to the limited availability of genome assemblies from species positioned between O. brachyantha and O. punctata in the phylogenetic tree, it has remained unclear whether the MBGC was restricted to species of the AA and BB lineages. Alternatively, the MBGC could have been lost specifically in the O. brachyantha lineage and might be present in other Oryza lineages. Lastly, the presence and configuration of the MBGC have not yet been studied in any of the allotetraploid Oryza species, and it therefore remains unknown if the cluster is present in both their sub-genomes.

Momilactone biosynthetic gene cluster in other species

Studies of the MBGC have not been limited to the Oryza genus; some have expanded into the wider Poaceae family. Poaceae encompass 12 subfamilies, of which nine belong to two core clades: BOP and PACMAD. The BOP clade comprises Bambusoideae, Oryzoideae (including Oryza), and Pooideae (including the Triticeae tribe); the PACMAD clade consists of Panicoideae, Aristidoideae, Chloridoideae, Micrairoideae, Arundinoideae, and Danthonioideae (Soreng et al., 2017). The MBGC is not restricted to the Oryza genus and has been identified through phylogenomic methods in species belonging to the PACMAD clade, specifically in the Panicoideae and Chloridoideae subfamilies (Wu et al., 2022a). However, it remains uncertain whether these species actually produce momilactones or other types of labdane-related diterpenoids. Echinochloa crus-galli (Panicoideae), a prevalent weed associated with rice cultivation, is the only species confirmed to produce momilactone A (Kraehmer et al., 2016; Wu et al., 2022a). Interestingly, E. crus-galli exhibits a distinct MBGC architecture, including an additional cytochrome P450 (EcCYP76L11) (Guo et al., 2017; Kitaoka et al., 2021). EcCYP76L11, similarly to CYP76M8 in Oryza sativa, can catalyse the conversion of 9βH-pimara-7,15-diene into 6β-hydroxy-syn-pimaradiene (Kitaoka et al., 2021). Despite the shared origin of the MBGC in Oryza and Echinochloa, the CYP76L11 gene has not been identified in the MBGC or the genome of any Oryza species (Wu et al., 2022a).

Here, we studied the presence, architecture, and evolution of the MBGC in cultivated rice and wild relatives from the Oryza genus by mining recently published genome assemblies. Our study shows that the MBGC is not restricted to species of the AA and BB lineages; we identify MBGC-like clusters in Oryza CC species and in one of the respective sub-genomes of the tetraploid species O. alta as well as in O. coarctata, a basal lineage in the Oryza phylogeny. We also show that the gene cluster was lost in several intermediate genome types. While momilactone A was detectable in all species harbouring the cluster, we were not able to detect momilactone B in O. coarctata, which might suggest that momilactone B production could have been a more recent innovation that emerged after the branching off of the KK genome type. In the O. coarctata MBGC, we identified an additional CYP that is different from the canonical O. sativa CYP99A2/3 and for which a corresponding orthologue could be found only in the MBGC of E. crus-galli, altogether suggesting the existence of an early ancestral cluster. In summary, our study shows how a biosynthetic gene cluster diversified within a plant genus while largely maintaining its biosynthetic function.